I need help debugging a random lockup

[John Griffin]

we have been working this issue long enough that we started noticing the eth
port and drive lights bumped at the same moment the system locked.
try this link for a driver update-
http://www.broadcom.com/drivers/downloaddrivers.php
I cannot find where I read this, but remember seeing that there were
improvements w/ this driver over the tigon3 driver. Does anyone know if the
native Redhat tigon3 drivers are better/worse suited for Broadcom 57xx than
the provided driver by Broadcom for Linux. I'm experiencing a learning curve
on kernel source tree and having issues implementing the driver linked
above.


----- Original Message ----- 
From: "Marshall Lewis" <marshall at novafoundry.com>
To: "AMD 64 Fedora List" <amd64-list at redhat.com>
Sent: Wednesday, May 05, 2004 5:02 PM
Subject: Re: I need help debugging a random lockup


> My workstation is actually running on a msi k8t master2-far.  Single
> opteron 246 w/1gig ecc ram (2 512 sticks).
>
> How'd you notice the activity on the eth port?  watching the light?
> I'm using the onboard sata, and it has an ide dvd-rom, but it's never
> been running during one of these lockups.
>
> I've noticed the some problems with the broadcom nic as well... problems
> with it being detected by the switch, sometimes I have to drop the
> interface and bring it backup before the switch will notice it.  I've
> also seen on some of our other opteron boxes (might be a coincidence..
> but I do believe the boxes that have had the problem are using the tg3
> driver) where the switch won't auto-sense the speed correctly (registers
> 100mbit instead of 1000mbit) but the nic thinks it's running at full
> speed.... obviously this leads to issues : )
>
> --
>  Marshall
>
>
> On Wed, 2004-05-05 at 16:27, John Griffin wrote:
> > we've been struggling w/ a lockup issue and I'm wondering if there is
any
> > relationship-
> > we have
> > dual opterons (244s)
> > MSI K8D Master-F.
> > ECC memory
> > we're having a devil of a time getting our CRON scripts to burn backups
on a
> > CDRW
> > I should say the isolated issue is on the portion of the script related
to
> > the CD Burn program- when the system is at the point of completing the
burn
> > (writing the table of contents?), our system locks up.
> > we noticed activity on our eth0 port and SCSI array at the moment the
system
> > locks.
> > our thought is there is a conflict w/ our RAID (through our Adaptec
> > 39320A-R).
> > we also have discovered issues in our driver for the Broadcom gig-e on
> > board. We have been unable to implement the updated drivers for the
> > Broadcoms and are still using the Tigon3 drivers.
> >
> > Has anyone had this type of lockup?
> >
> > John Griffin
> > V1 Datacom Inc
> >
> >
> >
> > ----- Original Message ----- 
> > From: "Marshall Lewis" <marshall at novafoundry.com>
> > To: <amd64-list at redhat.com>
> > Sent: Wednesday, May 05, 2004 1:10 PM
> > Subject: Re: I need help debugging a random lockup
> >
> >
> > > Well.. now I can reproduce it .. sorta..  of course right after
sending
> > > this email I decide to try and chart some stuff in oo calc, and it has
> > > locked 5 times in the last half hour.  I'm not any closer to
determining
> > > the cause though... maybe it's 32 bit apps? (or maybe it's power
> > > fluctuations, or the room is too hot, or my monitors are too close
> > > together... )        : )
> > >
> > > --
> > >  Marshall
> > >
> > > On Wed, 2004-05-05 at 15:34, Marshall Lewis wrote:
> > > > I'm not sure how to begin debugging this, it's not something that
I've
> > > > ever had happen on a linux system before.  I have an opteron
workstation
> > > > here at work, and it locks up more or less randomly... It's a
complete
> > > > system freeze.  There are no messages about it in /var/log/messages,
and
> > > > I'm guessing there aren't any messages on the consoles, although I
have
> > > > no real way to confirm.
> > > >
> > > > It doesn't seem to be application related (unless maybe it's X
itself or
> > > > the nvidia drivers), and is not load related.  .. it can happen
while
> > > > I'm clicking around in firefox, typeing on openoffice, coding in
nedit
> > > > or zend studio, highlighting text, etc.. etc... .. and of course it
even
> > > > happens when my back is turned (probably out of spite).
> > > >
> > > >  On any other machine I might suspect the ram.. and I guess I'm
kinda
> > > > suspecting the ram here, but it is an opteron, and it's running with
ECC
> > > > ram... so ram really shouldn't be an issue right?  or wrong? : )
> > > >
> > > >
> > > > Any suggestions on how I can determine the cause of the lockup? (or
at
> > > > least narrow it down).  BTW these lockups occured in FC1 as well
> > > > (currently running FC2T3).
> > > >
> > > > --
> > > >  Marshall
> > > >
> > >
> > >
> > > -- 
> > > amd64-list mailing list
> > > amd64-list at redhat.com
> > > https://www.redhat.com/mailman/listinfo/amd64-list
> >
>
>
> -- 
> amd64-list mailing list
> amd64-list at redhat.com
> https://www.redhat.com/mailman/listinfo/amd64-list